Figure 3.

Determining the quality of a step fit

(A) For every step fit the algorithm performs, the quality of the fit (orange line) is evaluated by means of an additional fit (blue line, called a counter fit). The counter fit is built by determining the next partition point (i_next), after which the current is rejected. Subsequently, the algorithm places the counter fit (blue) plateaus at locations within the existing fit (orange).

(B) Simulated trajectory representing a motor stepping behavior.

(C) Representative example of experimental and analytical S-curves obtained by fitting the trajectory in (B) through minimization of σ². Shaded areas indicate the underfitting (yellow) and overfitting (light blue) regime.

(D) Representative example of experimental and analytical S-curves obtained by fitting the trajectory in (B) through minimization of the sum of absolute differences (SAD²). Shaded areas indicate the underfitting (yellow) and overfitting (light blue) regime.

Also see Figure S2.